Etude 4 : Effet d’une solution hypertonique de lactate de sodium sur le

a. Justification

Suite à nos résultats sur l’œdème cérébral, l’hypoxie tissulaire et sur le versant pro apoptotique de la dysfonction mitochondriale post traumatique, nous avons voulu nous intéresser plus au volet métabolique. Nous avons donc souhaité analyser la dysfonction mitochondriale fonctionnelle versant phosphorylation oxydative et le métabolisme cérébral post traumatique avec étude du lactate endogène intra cellulaire.

Des études récentes se sont intéressées à l’apport de lactate de sodium exogène au cours d’un TC du fait de nombreuses publications soulignant son effet bénéfique endogène comme substrat énergétique alternatif en post traumatique en conditions de neuroglucopénie. Nous avons voulu appliquer les différentes techniques de monitorage précédentes à des rats traités par du lactate de sodium hypertonique afin de comprendre plus précisément les mécanismes d'action de cette molécule sur l’œdème cérébral, l’hypoxie tissulaire post traumatique, la dysfonction mitochondriale et le métabolisme cérébral.

Nous avons ainsi étudié les effets d'une administration de lactate de sodium exogène, a) sur l'œdème cérébral et l’hypoxie cérébrale post traumatiques et sur le métabolisme cérébral évalué en IRM de diffusion/BOLD et en spectroscopie protons respectivement b) sur le diamètre des vaisseaux étudiés en IRM c) sur les modifications mitochondriales morphologiques évaluées en microscopie électronique ainsi que d) sur la dysfonction de la chaine respiratoire mitochondriale évaluée par oxygraphie.

b. Hypothèses

Nos hypothèses étaient que :

- A la suite du TC diffus survient un œdème cérébral avec hypoxie tissulaire

cérébrale tout comme retrouvé dans nos deux études 1 et 2

- Une dysfonction mitochondriale fonctionnelle de la chaine respiratoire est

- Une accumulation de lactate intracellulaire se produit après notre TC expérimental du fait d’une hyperglycolyse ou d’une dysfonction mitochondriale

- Le lactate de sodium molaire de par ses propriétés hypertoniques limite

l’œdème cérébral et ainsi l’hypoxie tissulaire post traumatique

- Le lactate de sodium molaire en améliorant l’apport en O2 et en constituant un

substrat alternatif pour le cerveau lésé, améliore la dysfonction mitochondriale post traumatique de la chaine respiratoire.

- En améliorant la fonction de la chaine repsiratoire, il améliore ainsi l’utilisation

des substrats énergétiques disponibles et donc le métabolisme cérébral post TC en diminuant l’accumulation de lactate endogène utilisé pour le cerveau lésé.

- Par un effet vasodilatateur, le lactate de sodium améliore aussi le DSC et

donc l’apport en O2 aux tissus.

c. Méthodes spécifiques Thérapeutique étudiée

Dans l'étude 4, le lactate de sodium était injecté sous forme de débit continu dès 30 minutes après le TC à 0,5 ml/Kg/H sur 3H (1,5 mosm/Kg) chez les TC-lactate. Dans le groupe TC-saline, du NaCl 0,9% était injecté au même débit selon la même durée. Les groupes de rats non traumatisés recevaient les mêmes traitements.

Saturation tissulaire en O2, temps de transit moyen (TTM), volume sanguin cérébral (BVf) et diamère des vaisseaux (VSI) IRM

Cette étude était réalisée sur l’IRM 7T selon les mêmes modalités que l’étude 1. La mesure de la densité microvasculaire est réalisée par la mesure du Vessels Size Index (VSI). Cette mesure permet une analyse non invasive répétée globale du diamètre des vaisseaux. Elle permet l’analyse de la réactivité vasculaire dans des conditions normales ou pathologiques et permet d’attribuer à des thérapeutiques étudiées un effet de vasodilatation ou de vasoconstriction. Sa mesure implique une modification dans la relaxivité transverse qui est induite par l’injection intravasculaire d’un agent de contraste et l’acquisition d’une séquence multi gradient echo et spin-echo.

Spectroscopie protons

Dans l'étude 4, les séquences de SRM in vivo et d’IRM de diffusion pour objectiver l’œdème cérébral du modèle, ont été réalisées dans un aimant de 9,4T avant le TC (H0) puis répétées à H2 et H4 en post-TC. Les analyses spectroscopiques étaient réalisées afin d’évaluer les pics d’intensité de différentes molécules intervenant dans le métabolisme cérébral chez nos animaux et leurs variations après TC et administration de lactate, tels que le lactate, la créatine totale (TCr : somme créatine et phosphocréatine) et le N Acetyl Aspartate (NAA) (Figure 20). Cette mesure permet une analyse de ces molécules au niveau intracellulaire, contrairement aux techniques de microdialyse cérébrale, permettant d’interpréter les résultats obtenus et d’en déduire le comportement du métabolisme cérébral en lien avec l’analyse de la phosphorylation oxydative.

-

Figure 20: Spectre issu de l'analyse en spectroscopie H+ d'un rat TC-saline. TCr : Creatine et Phosphocréatine ; Cho : Choline ; NAA : N Acetyl Aspartate.

Chaine respiratoire mitochondriale

Après extraction mitochondriale, cette analyse était effectuée à l'aide d'une électrode de Clark dans un milieu contrôlé en continu sur le plan thermique circulant au sein d'une chambre contenant un tampon de respiration. Après équilibration, pour tester les complexes I-III-IV et II-III-IV, le Glutamate Malate ou le Succinate, substrats des complexes I et II respectivement, ont été ajoutés. L'ADP était alors introduit pour

initier l'état III pendant 2 minutes (ou stade de phosphorylation oxydative), suivi par l'addition d'oligomycine pour monitorer l'état IV de la respiration pendant deux autres minutes. Le rapport état III/ état IV correspond au respiratory control ratio (RCR) et sert à évaluer la fonctionnalité de la chaine respiratoire. Cette mesure tout comme la mesure de la CRC, sont des mesures in vitro des capacités maximales de la chaine respiratoire des mitochondries totales de l’animal testé. Il ne s’agit pas d’un reflet exact de la fonctionnalité de la chaine repsiratoire dans son milieu in vivo. Cependant, la comparaison des résultats du RCR des différents groupes de rats permet de conclure quant à l’altération de la chaine respiratoire dans les différentes conditions testées.

Dans ce travail, j'ai réalisé toutes les manipulations IRM et spectroscopiques, avec les aides de Vasile Stupar concernant la mise en place de la technique 1H-MRS sur la plateforme IRM du GIN et de Florence Fauvelle concernant l'interprétation des données des spectres de spectroscopie. J’ai encadré Adrien Cuisinier dans la réalisation de ses manipulations pour la partie hypoxie cérébrale et analyse du VSI et dans ses analyses statistiques. Ce travail a également été fait en collaboration avec le LBFA et Cécile Batandier concernant l'analyse de la chaine respiratoire des mitochondries. Benjamin Lemasson nous a aidés dans l’analyse des données IRM. J’ai ensuite effectué l’analyse statistique des données et rédigé l’article avec l’aide du Pr Payen. Nora Collomb nous a aidés dans la réalisation des manipulations. L’analyse en microscopie électronique à été réalisée grâce à l’aide de Karin Pernet Gallay et d’Anne Bertrand.

Hypertonic sodium lactate to modulate brain metabolism

dysfunction following diffuse traumatic brain injury

Anne MILLET, M.D. ^1,2,3 Adrien CUISINIER, M.D. ^1,2,4 Pierre BOUZAT, M.D., PhD. ^1,2,4 Cécile BATANDIER, Ph.D. ⁵ Benjamin LEMASSON, Ph.D. ^1,2 Vasile STUPAR, Ph.D. ^1,2 Karin PERNET-GALLAY, Ph.D. ^1,2 Emmanuel L. BARBIER, Ph.D. ^1,2

Jean François PAYEN, M.D., Ph.D. ^1,2,4

1

INSERM, U1216, F-38000, Grenoble, France

2

Univ. Grenoble Alpes, Grenoble Institut des Neurosciences, F-38000 Grenoble, France

3

Pôle Couple Enfant, Hôpital Michallon, CHU Grenoble Alpes, F-38000, Grenoble, France

4

Pôle Anesthésie Réanimation, Hôpital Michallon, CHU Grenoble Alpes, F-38000, Grenoble, France

5

INSERM, U1055, Laboratoire de Biologie Fondamentale et Appliquée, Grenoble, F-38000, France

Correspondence: Dr Jean-François Payen, Grenoble Institut des Neurosciences, INSERM

U1216/Equipe 5. Université Grenoble Alpes, Site santé de la Tronche, BP 170. Grenoble Cedex 9, 38042 France. Tel 33 4 56 52 05 89. Fax 33 4 56 52 05 98. Email: jfpayen@ujf-grenoble.fr

Abstract

Background: Hypertonic sodium lactate (HSL) solution is emerging as a promising option to treat injured brain tissue. However, its exact mechanism of action remains unclear. Using multiparametric MRI-based approaches, we investigated the cerebral metabolic and circulatory effects of a post-traumatic infusion of HSL in a rodent model of diffuse traumatic brain injury (TBI).

Methods: Thirty minutes after trauma (impact-acceleration model), adult male Wistar rats were randomly assigned to receive a 3 h intravenous infusion of either a saline solution (TBI-saline) or hypertonic sodium lactate (TBI-HSL). Two other groups received no trauma insult (sham-saline and sham-HSL). Three series of experiments were conducted up to 4 h (H4) after TBI. The first investigated the effects of HSL on brain oedema using diffusion-weighted MRI to measure apparent diffusion coefficient, and on brain metabolism using in vivo localised ¹H MR spectroscopy to measure intracellular lactate/creatine ratio (n = 10 rats per

group). The respiratory control ratio was then determined at H4 using oxygraphic analysis of extracted mitochondria from brain tissue. Secondly, the effects of HSL on brain oxygenation and perfusion were investigated using a multiparametric quantitative blood oxygenation level–dependent MR approach (n=10 rats per group) to measure tissue oxygen saturation (StO2) and vessel size index. Finally, mitochondrial ultrastructural changes after treatment

were studied (n = 1 rat per group).

Results: Compared with the TBI-saline group, the TBI-HSL and the sham-operated groups had reduced brain oedema (P<0.05). Concomitantly, the TBI-HSL group had significantly lower intracellular lactate/creatine ratio, higher mitochondrial respiratory control ratio,

improved StO2 and vessel size index measurements, and reduced morphological

mitochondrial disruption in astrocytes and neurons compared with the TBI-saline group (all P<0.05).

Conclusion: These findings indicate that post-traumatic administration of HSL can reverse brain oxygenation and metabolism dysfunction after diffuse TBI. HSL has versatile properties that make this agent particularly suitable to reverse brain tissue derangements following brain trauma.

List of abbreviations:

ADC: apparent diffusion coefficient

BOLD: blood oxygenation level–dependent BSA: bovine serum albumin

1

H-MRS: ¹H-magnetic resonance spectroscopy

Hb: haemoglobin content HSL: hypertonic sodium lactate ICP: intracranial pressure Lac: intracellular lactate

MABP: mean arterial blood pressure NAA: N-acetyl-aspartate

RCR: respiratory control ratio TBI: traumatic brain injury TCr: total creatine

TE: echo time TR: repetition time VSI: vessel size index

Introduction

There has been a growing interest in the exogenous supplementation of lactate through the administration of HSL solution following brain injury (Bouzat and Oddo, 2014). The rationale for using such a solution was based on accumulating evidence of endogenous lactate as a major oxidative substrate utilised by injured brain cells (Berthet, 2009; Gallagher, 2009; Schurr, 1988). It was shown that exogenous lactate anions could also enter injured brain cells to be utilised as an energy source (Chen, 2000; Maran, 1994; Meierhans, 2012). In patients with no evidence of brain ischemia after TBI, HSL was associated with a significant increase in extracellular concentrations of lactate, pyruvate and glucose (Bouzat, 2014). It was suggested that HSL would act as a glucose-sparing substrate through increasing pyruvate availability for the mitochondrial tricarboxylic acid cycle, restoring oxidative metabolism after TBI. Because a failure of mitochondrial respiration was described after TBI (Lifshitz, 2004; Singh, 2006), HSL could ultimately reverse such metabolic dysfunction. This implies that mitochondrial function in injured brain tissue will benefit from such lactate flooding, a hypothesis that remains unexplored (Dienel, 2016). Alternatively, the cerebral effects of HSL may be due to lactate-related vasodilating effects (Alessandri, 2012; Gordon, 2008) and/or hyperosmotic and anti-oedematous effects leading to a reduction in ICP (Bouzat et al., 2014; Ichai, 2009; Ichai, 2013).

In order to clarify the exact role of HSL in brain injury, we used multiple techniques to determine intracellular lactate, brain tissue oxygenation and perfusion, and mitochondrial RCR in a rodent model of diffuse TBI. We hypothesised that exogenous lactate would be utilised by injured brain cells to reverse brain metabolism and oxygenation dysfunction induced by diffuse TBI. This assumption implies versatile effects of HSL on both brain perfusion and mitochondrial respiration.

Materials and Methods

A series of three experiments were conducted on each of four groups of adult male Wistar rats (350-500g). In the first series of experiments, we studied the effect of HSL on combined measurements of brain oedema and brain metabolism after diffuse TBI. We used diffusion-weighted MRI to assess brain oedema prior to TBI, at 2 h (H2) and 4 h (H4) after TBI or equivalent time, thus each rat (n = 10 per group) acted as its own control. Simultaneously, brain metabolism was assessed using in vivo localised ¹H-MRS. Rats were

then sacrificed to measure the RCR of extracted mitochondria from brain tissue using oxygraphic analysis. Rats were randomly assigned to one of four groups: the TBI-saline group received a saline treatment and the TBI-HSL group was treated 30 min post-trauma for 3 h with HSL. Sham-saline and sham-HSL rats received no TBI insult. In the second series of experiments using a similar protocol and four groups of rats, we investigated the effect of HSL on combined measurements of brain perfusion and brain oxygenation. We used multiparametric quantitative BOLD approach to measure brain tissue oxygen saturation (StO₂) (n = 10 per group) and VSI (n = 5 rats per group) prior to TBI and at H2 and H4 after

TBI or equivalent time. In the third series of experiments using a similar protocol, we evaluated morphological disruption of mitochondria using electron microscopy at H4 (n = 1 rat per group).

* Experimental protocol

The study design was approved by the local Internal Evaluation Committee for Animal Welfare and Rights. Experiments were performed in accordance with the guidelines of the French Government (licenses 380819 and A3851610008). The experimental protocol was similar to the one described previously (Bouzat, 2013; Millet, 2016). In the animal facilities, rats were maintained within a 12 h light-dark cycle with humidity and temperature controlled

at normal levels, and allowed food and water ad libitum (2-4 cage companions). Anaesthesia was induced and maintained using isoflurane (inhaled fraction of 2.0-2.5%) during the study period. After tracheal intubation, rats were mechanically ventilated with 60% air–40% oxygen using a rodent ventilator (SAR-830/P, CWE, Ardmore, PA, USA). Catheters were inserted

into the femoral artery and vein for MABP monitoring and drug delivery. Blood gases (PaO2

and PaCO2), arterial pH and Hb were determined from arterial blood samples of less than 0.1

ml (ABL 510, Radiometer, Copenhagen, Denmark). Rectal temperature was kept at 36±0.5 °C. Criteria for exclusion from the study were values outside the range of the following physiological parameters measured 1 h (H1) after the reference time: MABP < 70 mm Hg, PaO2 < 100 mm Hg, PaCO2 < 30 mm Hg or > 50 mm Hg, and Hb < 80 g/L.

Injury was induced by dropping a 500 g mass through a vertical cylinder from a height of 1.5 m on to a metallic disc glued to the central area of the skull vault between the coronal and lambdoid sutures, according to the initial description of impact-acceleration model. The reference time (H0) corresponded to the moment of impact (TBI-saline and TBI-SL) or to an equivalent time (sham-operated). Thirty minutes after H0, the rats were randomly allocated to intravenously receive either 0.5 ml/kg/h of an isotonic saline solution or 0.5 ml/kg/h of a

solution of 11.2% HSL (Na⁺ 2000 mmol/L, lactate 2000 mmol/L, prepared by the Assistance

Publique des Hopitaux de Paris, France) during a continuous infusion lasting 3 h. Randomisation was achieved using pieces of paper folded and placed in a receptacle, with each piece of paper recording an animal number. A piece of paper was randomly withdrawn without replacement on the day of each experiment to allocate each animal into a group.

* MRI measurements

In the first series of experiments, MRI was performed before, 2 h (H2) and 4 h (H4) after TBI or an equivalent time in sham-operated groups at 9.4T in a horizontal bore magnet

(Biospec Avance III HD, Bruker Biospin, Ettlingen, Germany). A 72 mm inner diameter quadrature volume coil was used for radio frequency transmission and an actively decoupled four channel array surface coil optimised for rat brain MRI was used for signal detection. A

T2-weighted TurboRARE sequence (TR/TE = 2500 ms/33 ms, voxel size = 117x117x800 µm,

23 slices, total acquisition time of 2 min 40 s) was used to delineate the region of interest. Diffusion-weighted images were acquired using a diffusion-weighted MRI protocol. Five adjacent rostrocaudal slices were acquired using a 2D diffusion-weighted, spin-echo, single-shot echo-planar imaging (TR 2000 ms, TE 20 ms, field of view 25x25 mm, matrix 108x108, 8 averages; voxel size 278x278x800 µm, total acquisition time of 2 min 24 s). This sequence was applied six times, three without diffusion weighting and three times with diffusion weighting (b = 650 s.mm^-1) in three orthogonal directions, in order to map the ADC (mm²/s).

One region of interest was manually drawn on the T2-weighted images and copied onto the ADC images, including right and left parietal neocortex, and right and left caudoputamen. The change in ADC was measured at H2 (=ADC H2-ADC before TBI) and at H4 (=ADC H4-ADC before TBI).

In vivo ¹H-MRS was acquired at the same time points. Spectra were acquired with the

short echo time PRESS sequence (Bottomley, 1987), with TE/TR=16.16/2500 ms, 4401 Hz bandwidth, 2048 data points and 256 averages. The total acquisition time was 11 min. The volume of interest was set to 27 mm³ (3x3x3 mm) and placed in the centre of the left striatum.

The PRESS localisation was preceded with the VAPOR water suppression and outer volume suppression modules. Spectra were quantified using the subtract-QUEST procedure (Ratiney, 2005) running under jMRUI package software (Java-based version of the Magnetic Resonance User Interface, http://www.mrui.uab.es/mrui). The value of intracellular NAA and Lac amplitudes was normalised to the sum of creatine and phosphocreatine that is total TCr.

In the second series of experiments, MRI was performed at 7T in a horizontal bore magnet (Bruker Biospec 47/40 USR AV III, Bruker BioSpin, Wissembourg, France). StO₂

maps were obtained using a multiparametric quantitative BOLD approach as we previously described (Christen, 2011; Christen, 2014). Haematocrit was fixed to 0.85 × 42%. The 0.85 factor accounts for the Fåhræus-Lindqvist effect that reduces the haematocrit in small capillaries. Cerebral perfusion was measured using a previously described steady-state approach based on the change in susceptibility after injection of ultrasmall superparamagnetic particles of iron oxide (P904, 200 µmol/kg, Guerbet SA, Aulnay-sous-Bois, France) which allows the determination of VSI (Tropres, 2001). The VSI changes ( VSI) were calculated at H2 and at H4 by comparison with H0: VSI = (VSI at H2 (or H4) – VSI at H0)/VSI at H0. Additionally, serum concentrations of lactate, sodium, chloride and osmolality were determined at H4, i.e., 1 h after the cessation of infusion.

* Mitochondria extraction and oxygraphic analysis

Once ¹H-MRS was completed at H4 (first series of experiments), anaesthetised

animals were sacrificed by decapitation and brains were rapidly excised. The procedure was carried out at 4°C. Once removal of cerebellum and meninges, brain tissue was placed in a Potter Elvejhem homogeniser containing MSHE buffer (225mM mannitol, 75 mM sucrose, 10 mM HEPES, 1 mM EGTA, pH 7.40) and crushed at 500 rounds per minute in the presence of BSA (fatty acid free, 0.1%). After ten strokes, the homogenate underwent four successive differential centrifugations. The first one was conducted at 2000 g for 4 min to remove cellular debris and nuclei. The supernatant was centrifuged at 12000 g for 10 min, carefully removed and the pellet was resuspended in MSHE-BSA buffer and incubated with 0.02% digitonin during 4 min for permeabilising synaptosomes. The pellet was then subjected to two final centrifugations at 12000 g for 10 min each and resuspended in buffer isolation. The

protein concentration of the suspension was assessed at the beginning of each procedure using the Bradford technique (Hammond and Kruger, 1988). The yield of mitochondria was set in 0.5 mg/ml suspension buffer for each animal. Once extracted, fresh mitochondria were used within 4 h.

Oxygraphic analysis was then used to assess mitochondrial RCR. The protocol was similar to the one previously described (Millet et al., 2016). We used a Clark-type electrode (StrathKelvin, Scotland) at 37°C in a continuously stirred, sealed thermostatically controlled chamber. Isolated mitochondria (0.5mg/ml) were placed in the chamber containing respiration KET buffer (125 mM KCl, 1 mM EGTA, 20 mM Tris, pH 7.2) in the presence of 5mM phosphate and BSA and tested with different inducers. After equilibration, glutamate-malate (5 mM/2.5 mM) or succinate (5 mM), which are substrates of complexes I and II respectively, were added. Adenosine diphosphate (300µM) was then added to initiate a state-III respiratory rate for 2 min, followed by the addition of oligomycin (1µg/ml) to obtain a non-phosphorylating state for 2 min. The RCR was calculated dividing the state III respiratory rates by non-phosphorylating state rates.

* Electron microscopy

In the third series of experiments using a similar protocol, rats were sacrificed by decapitation at H4 (n = 1 rat per group). Brains were rapidly excised and cortex samples (n = 3 per rat) fixed in 2.5% glutaraldehyde (0.1 M cacodylate buffer, pH 7.4) for 24 h at room temperature. Samples were post-fixed in 1% osmium tetroxide (0.1 M cacodylate, pH 7.2) for 1 h at 4°C, and stained with 1% uranyl acetate (pH 4) in water for 1 h at 4°C and dehydrated through graded alcohol concentrations before being embedded in Epoxy Embedding Medium (Fluka, Sigma-Aldrich, Saint-Quentin Fallavier, France). Cortex brain ultrathin sections (60 nm-thickness) were stained with 5% uranyl acetate and 0.4% lead citrate before being

observed under a transmission electron microscope at 80 kV (Jeol 1200 EX, JEOL Ltd, Tokyo, Japan). Images were acquired with a digital camera (Veleta, SIS, Olympus, Munster, Germany). The analysis was performed by an examiner blinded to the treatment allocated to